US 12,032,544 B1
Systems, methods, and computer products for model-based query assignment
Michael G. Bradley, McLean, VA (US); James Douglas Gordon, McLean, VA (US); Stanislav Kornienko, McLean, VA (US); and Gary Skinner, McLean, VA (US)
Assigned to FEDERAL HOME LOAN MORTGAGE CORPORATION (FREDDIE MAC), McLean, VA (US)
Filed by Federal Home Loan Mortgage Corporation, McLean, VA (US)
Filed on Mar. 7, 2022, as Appl. No. 17/687,948.
Application 17/687,948 is a continuation of application No. 16/050,579, filed on Jul. 31, 2018, granted, now 11,269,827.
Claims priority of provisional application 62/539,132, filed on Jul. 31, 2017.
Int. Cl. G06F 16/21 (2019.01); G06N 5/00 (2023.01); G06Q 20/40 (2012.01); G06F 16/215 (2019.01); G06F 16/903 (2019.01); G06Q 40/03 (2023.01)
CPC G06F 16/219 (2019.01) [G06F 16/215 (2019.01); G06F 16/903 (2019.01); G06N 5/00 (2013.01); G06Q 20/4016 (2013.01); G06Q 40/03 (2023.01)] 14 Claims
OG exemplary drawing
 
1. A system comprising:
a database;
a processor in communication with the database; and
a memory device, in communication with the processor, storing instructions that perform operations comprising:
maintaining at the database a data structure comprising historical fraud data by:
retrieving historical fraud data from a first data source; and
generating aggregated historical fraud data by aggregating the historical fraud data from the first data source and historical fraud data from a second data source;
receiving a query requesting a streamlined process based on the query, instantiating a fraud model by:
obtaining the aggregated historical fraud data from the database and determining, based on the aggregated historical fraud data, a modeled fraud ratio,
determining, based on the aggregated historical fraud data, a submitted fraud ratio,
making a first assessment whether the submitted fraud ratio falls below a given threshold based on an assessed connection to fraud, and
based on the first assessment, assigning a first flag to the query;
further based on the query, instantiating a condition model by:
determining a difference between the fraud ratio and the given threshold, and
based on the determined difference, assigning a second flag to the query; and
based on the first flag and the second flag, determining whether to assign the query to the streamlined process.